Quantification of Croatian phonological and morphological noun categories in the Bible

U radu se izdvojene imenice iz cijele građe Biblije u hrvatskomu prijevodu na temelju Velike konkordancije raščlanjuju na nekoliko gramatičkih kategorija. Na početku se opširnije predstavlja građa i način kako se pristupilo izboru natuknica. Potom se raščlanjuju fonološke kategorije: završetak imenica i nepostojano a, koje se razgraničuje od slične pojave umetanja -a- u genitivu množine. Morfološke se kategorije najprije prikazuju odvojeno: sklonidba, rod, broj i živost. Na kraju se u međuodnos stavljaju završetak imenice, sklonidba i rod. Pokazuje se da je unutar svih imeničkih natuknica samo pet složenih kategorija zastupljeno više od pet posto: najviše imenice muškoga roda na -K koje pripadaju a-sklonidbi (npr. sin, kralj), potom imenice ženskoga roda na -a koje pripadaju e-sklonidbi (npr. ruka, žena), imenice srednjega roda na -e koje pripadaju a-sklonidbi (npr. ime, more), imenice ženskoga roda na -K koje pripadaju i-sklonidbi (npr. riječ, krv) i imenice srednjega roda na -o koje pripadaju a-sklonidbi (npr. djelo, nebo). Sve ostale kategorije imaju udio manji od jedan posto. Iako je očito da neke pojedinosti ovise o polaznoj građi, pretpostavlja se da opći odnosi u velikim promatranim imeničkim kategorijama mogu poslužiti kao pokazatelj zastupljenosti tih kategorija u hrvatskim tekstovima općenito. Slika fonoloških i morfoloških obilježja hrvatskih imenica u ovladavanju hrvatskim izvornojezičnih i inojezičnih učenika na temelju dobivenih podataka može biti jasnija ako se najprije daju obrisi najzastupljenijih i svojstvenih članova pojedinih kategorija, a tek poslije rjeđe, manje i nesvojstvene koji pripadaju različitim kategorijama, posebno ako su one isprepletene.The paper analyzes several grammatical noun categories in Croatian Bible translation. The data is collected from the Bible Concordance (Vojnović 1991), based on the Old Testament translation as found in the Zagreb Bible and the New Testament translation of Duda, Fućak (1973, 1991). The nouns are lemmatised and the number of tokens are assigned to each lemma, but proper nouns are excluded. Two phonological categories are analyzed: the word ending (closed or open syllable) and movable a (the nucleus of the final closed syllable which breaks consonant cluster), which is distinguished from Genitive plural a insertion, as the second is morphologically based. There are 48% lemmas with closed syllables, 31% lemmas ending in -a, 6% in -o, 15% in -e and 0% in -i, while no lemma ends in -u or -ie. Among all lemmas 8% end in movable a, but among only -K and -o (vocalized -l) nouns, 18% end in movable a. Four morphological categories are analyzed: declension type, gender, number and liveliness. A-declension takes 60%, E-declension 31% and I-declension 9% lemmas. Masculine gender takes 41%, feminine 39% and neuter 20% lemmas. As only 21 nouns belong to the class of pluralia tantum, 100% of lemmas are singular. Among all lemmas 17% are marked for liveliness, but 41% among masculine nouns with -K or -o (vocalized -l) endings. The third part of the paper calculates the word ending, declension type and gender interface as they are highly predictable. It is found that only five complex groups have more than five percent in the data: 1. masculine nouns ending in -K which belong to A-declension: 40% (the five most frequent are: sin `son\u27, kralj `king\u27, narod `people\u27, Bog `God\u27, otac `father\u27, čovjek `man\u27), 2. feminine nouns ending in -a which belong to E-declension: 30% (e.g. ruka `hand\u27, zemlja `earth, land\u27, žena `woman, wife\u27, godina `year\u27, kuća `house\u27), 3. neuter nouns ending in -e which belong to A-declension (srce `heart\u27, ime `name\u27, dijete `child\u27, lice `face\u27, more `sea\u27): 14%, 4. feminine nouns ending in -K which have their own I-declension: 8% (riječ `word\u27, krv `blood\u27, zapovijed `commandment, order\u27, smrt `death\u27, mudrost `wisdom\u27), 5. neuter nouns ending in -o which belong to A-declension (oko `eye\u27, mjesto `place\u27, djelo `deed\u27, tijelo `body\u27, nebo `heaven, sky\u27): 5%. All other combinations have less than 1%. The last part of the paper lists many generalizations about the relationships between endings, declension-types and gender where quantifiers are supported by percentage: always (100%), almost always (97-99,99%), large majority (91-95,56%), majority (66-83%), minority, lesser part (10-33%), rarely (1-4%), very rarely (less than 1%). As some of the frequencies are in line with previous findings from different types of texts, it is believed that the frequency of each analyzed category is not typical for the Bible only, but for Croatian language in general. Hence, when only the most typical combinations are considered, the picture of Croatian morpho(phono)logy of nouns does not seem to be as complicated as when all combinations are taken into account. The above numbers may be used to support approaches in teaching and learning Croatian as L2 which rely on prototypicality when beginners are concerned